I wanted to provide a preview of some updates on the Foundation we’ll be talking about in Sydney. This email is a little long, so here’s the quick summary—read on for more detail.
We’re seeing some common issues that open infrastructure users are running into when deploying OpenStack and other open source infrastructure projects together, and we want to improve this situation through Foundation efforts focused on the overall state of integration and operations for these users.
1) Planning to reorganize some Foundation community activities more around use cases such as datacenter cloud infrastructure (traditional OpenStack use case), CI/CD (very common use case), container infrastructure (rapidly growing use case), and edge infrastructure (emerging use case)
2) Supporting additional user-driven OSS projects to facilitate integration in these use case (not necessarily OpenStack-branded or part of the existing OpenStack datacenter infrastructure cloud services)
3) Expanding ongoing testing of OpenStack + other open source cloud tools (e.g. OpenStack in CNCF CI, new "OpenLab" testing effort, encouraging broader use of Zuul v3)
Now the longer version. As some of you may know a joint leadership group made up of the TC, the UC, the Board and Foundation staff has been doing some work this year to look at where we are 5 years in and what we should be thinking about over the next few years.
As we’ve talked about a lot of different aspects of OpenStack and also talked with users, we’ve seen some common trends emerge. First, we continue to see a lot of growth in OpenStack deployments. Second, we also see increasing diversity in the kinds of use cases organizations are deploying OpenStack to serve. We’ve also seen that many users are having difficulty with consuming all of the OpenStack services along with the numerous other open technologies that they often combine. While some of these issues are being helped by more mature commercial offerings in the market, in some cases there are areas that present us with opportunities as a community of developers and users working together.
One of the biggest needs is for better integration between open infrastructure components up and down the stack, and we have identified four key elements to improve the state of integration and operations by directing Foundation efforts and activities along these lines.
JOINT USE CASES
For a successful integration, we need user-driven reasons to combine components. We have hundreds of reasons from users all over the world. Organizations like Yahoo! Japan, eBay, AT&T, Liveperson, Commonwealth, Fico, GAP, Target, JD.com, SKT, GE Healthcare, SAP, China Railway, Fiberhome, Walmart, and China Telecom all combine OpenStack software with other open source tools including Kubernetes, Ceph and Cloud Foundry to solve real business problems.
To better define the use cases, we’re planning to organize some of our activities across problem domains:
• Datacenter cloud infrastructure
• Container infrastructure
• Edge infrastructure
• CI/CD infrastructure
• In the future other drivers of infrastructure usage like machine learning and AI
The OpenStack project has always been about enabling truly open collaboration across the industry to produce open source infrastructure software. Tackling these use cases more directly brings our open collaboration model to new areas and new organizations. It will also bring in new users and resources that will benefit our core cloud infrastructure components, users and resources that may not have previously seen where their needs fit in with one single giant thing called "OpenStack."
We have to get the communities together, build trust and deliver standard interfaces we can count on between the components. Interoperability efforts are key to ensuring that these integrations are repeatable, stable and production quality. Within the OpenStack community, we have many upstream contributors to other projects, and often organize events that bring together the developers and users across the communities.
FOSTERING NEW OPEN INFRASTRUCTURE PROJECTS
The rate of open source innovation has skyrocketed in the last 5 years, however, many projects are developed independently of the production environments in which they must run. We need new projects that focus on the practical implications of deploying large open source software projects together and operating them reliably for years. Lifecycle management of OpenStack has been a top need for users and projects like TripleO, Kolla, OpenStack-Ansible and the new openstack-helm project are tackling this problem head on.
When you look outside of the OpenStack core, you see issues such as container security that are preventing adoption among regulated industries. Edge computing and edge clouds represent an incredible opportunity, but have specific deployment and security needs that may require specialized components. These technology components are critical to deployment success and are often being developed separately inside of user organizations. We can help these users open source and collaborate together to build these systems even though some may not necessarily be OpenStack-branded or part of the existing OpenStack datacenter infrastructure cloud services.
TESTING OF END-TO-END SYSTEMS
Once we have identified a valid use case, ramped up the collaboration and built the tools that make the integration a reality, we need to test the entire end-to-end stack to ensure stability and performance. Many open source communities have excellent internal continuous testing, but users do not run our software in silos, and we can’t test in silos. The OPNFV XCI cross-community CI system is an example that implements this concept today. The new OpenLab initiative from OpenStack community members takes this concept to even more of these projects.
The OpenStack Foundation manages a large and growing global community of developers and users who are building and deploying open source software for managing infrastructure in all forms. This is a really exciting mission to be on right now, and I think that updating our strategy to support more integration and operation work is key to helping our users succeed not just with OpenStack clouds, but with their broader open infrastructure ambitions.
For those of you who will be in Sydney, we'll be talking more about this and we have a Forum session set up as well on Wednesday in room C4.11: https://www.openstack.org/summit/sydney-2017/summit-schedule/events/20597/openstack-foundation-update
Foundation mailing list